Post-traumatic Stress Symptoms Following Childbirth
Post-traumatic Stress Symptoms Following Childbirth
Mean maternal age at delivery was 31.2 years (SD 4.6 years, range 18.8–45.4 years); 5.2% of the women had previously delivered by a CS, 18.4% had a medical risk for an elective CS (including previous CS) and 47.2% were first-time mothers. Most of the women were married or cohabitating (97.8%) and did not smoke at the time of delivery (96.1%). Sixty-eight percent had more than 12 years of education. Compared with the national data from the Medical Birth Registry of Norway from 2009, the women in the study were less likely to be smokers (3.9% vs. 8.2% at the time of delivery) and, were slightly older (mean age of 31.2 years vs. 29.7 years), and there were fewer single women in the study (2.3% vs. 9.1%).
Table 1 shows the means and SDs for all variables. In our sample, 6.0% had scores above 19 on the Impact of Event Scale indicating clinically significant levels of stress, and 1.8% scored above 34, indicating that a PTSD condition was likely. The average score for post-traumatic stress symptoms following childbirth was 6.70 (SD = 7.99) ( Table 1 ). The mean score for the subscale intrusion was 4.28 (SD = 4.85) and for the subscale avoidance 2.35 (SD = 3.89). Of the women in the sample, 9.7% would have preferred to have a CS, and 5.6% actually delivered by elective CS. Most women (N = 1,493; 87.8%) belonged to the Match 1 group (no preference for CS, no elective CS). Match 2 (preference for CS, elective CS) comprised 53 women (3.1%). Mismatch 1 (no preference for CS, elective CS) and Mismatch 2 (preference for CS, no elective CS) comprised 42 (2.5%) and 112 (6.6%) women, respectively. The four match/mismatch groups differed substantially in terms of their medical risk (F = 93.16, p < 0.001) and fear of childbirth (F = 37.82, p < 0.001) ( Table 1 ). Fear of childbirth was particularly high in Match 2 and Mismatch 2, groups comprising women who had a preference for CS, whereas medical risk was particularly high in Match 2 and Mismatch 1, comprising women who actually delivered by elective CS.
Regarding post-traumatic stress symptoms, we found a significant interaction effect between preference and actual mode of delivery (F = 7.15, p = .008), and in examining differences for all four match/mismatch groups, ANOVA yielded significant overall group differences (F = 11.96, p < 0.001). However, although the Mismatch 1 and Mismatch 2 groups had a tendency toward higher levels of post-traumatic stress symptoms than the Match 1 and Match 2 groups, Bonferroni post-hoc tests found significant differences only between Match 1 (no preference for CS, no elective CS) and Mismatch 2 (preference for CS, no elective CS) ( Table 1 ).
The correlations with putative risk factors showed that post-traumatic stress symptoms were most strongly related to symptoms of depression and anxiety, but fear of childbirth and neuroticism also showed considerable correlations with such symptoms ( Table 2 ). When risk factors were included one by one as covariates in ANCOVA, the inclusion of fear of childbirth had the greatest potency in explaining group differences in post-traumatic stress symptoms, as the F-value of the group effect was reduced from 11.96 to 4.29. Symptoms of depression and anxiety (reduced to 7.35 and 7.61), and to some degree neuroticism and prior PTSD (reduced to 8.46 and 8.79), also decreased the F-value. However, even though the F-value was substantially reduced by some of the covariates, the overall group effect and the difference in the level of post-traumatic stress symptoms between Match 1 and Mismatch 2 still remained significant, as indicated by post hoc tests. When all risk factors were entered blockwise, all risk factors remained statistically significant ( Table 3 ). However, the block with the psychological risk factors showed the greatest potency in explaining group differences, compared to personality variables and somatic and demographic factors. Including the psychological factors reduced the F-value to 3.17, but the group differences remained significant ( Table 3 ). Within the block with psychological risk factors, fear of childbirth was the most important covariate (F = 32.76,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.019), followed by symptoms of depression (F = 26.05,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.015) and anxiety (F = 22.77,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.013). Within the block comprising personality, neuroticism was the most important covariate (F = 78.73,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.044), and within the block with somatic and demographic factors, parity had the largest effect (F = 20.41,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.012) ( Table 3 ).
When all covariates were entered simultaneously in one ANCOVA (they were all statistically significant in the previous model and thus included), the F-value was further reduced to 2.66; however, the group differences remained significant ( Table 3 ). In this model, symptoms of depression (F = 26.55,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.016), parity (F = 22.13,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.013), fear of childbirth (F = 20.85,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.012) and symptoms of anxiety (F = 17.74,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.010) were the most important covariates. The other risk factors (except prior PTSD, F = 3.98,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.002) were no longer statistically significant ( Table 3 ).
All ANCOVA were, furthermore, repeated with log transformed scores of the Impact of Event Scale. Besides minor changes in the F-values, the results were very similar to the initial analyses without transformation.
Results
Mean maternal age at delivery was 31.2 years (SD 4.6 years, range 18.8–45.4 years); 5.2% of the women had previously delivered by a CS, 18.4% had a medical risk for an elective CS (including previous CS) and 47.2% were first-time mothers. Most of the women were married or cohabitating (97.8%) and did not smoke at the time of delivery (96.1%). Sixty-eight percent had more than 12 years of education. Compared with the national data from the Medical Birth Registry of Norway from 2009, the women in the study were less likely to be smokers (3.9% vs. 8.2% at the time of delivery) and, were slightly older (mean age of 31.2 years vs. 29.7 years), and there were fewer single women in the study (2.3% vs. 9.1%).
Table 1 shows the means and SDs for all variables. In our sample, 6.0% had scores above 19 on the Impact of Event Scale indicating clinically significant levels of stress, and 1.8% scored above 34, indicating that a PTSD condition was likely. The average score for post-traumatic stress symptoms following childbirth was 6.70 (SD = 7.99) ( Table 1 ). The mean score for the subscale intrusion was 4.28 (SD = 4.85) and for the subscale avoidance 2.35 (SD = 3.89). Of the women in the sample, 9.7% would have preferred to have a CS, and 5.6% actually delivered by elective CS. Most women (N = 1,493; 87.8%) belonged to the Match 1 group (no preference for CS, no elective CS). Match 2 (preference for CS, elective CS) comprised 53 women (3.1%). Mismatch 1 (no preference for CS, elective CS) and Mismatch 2 (preference for CS, no elective CS) comprised 42 (2.5%) and 112 (6.6%) women, respectively. The four match/mismatch groups differed substantially in terms of their medical risk (F = 93.16, p < 0.001) and fear of childbirth (F = 37.82, p < 0.001) ( Table 1 ). Fear of childbirth was particularly high in Match 2 and Mismatch 2, groups comprising women who had a preference for CS, whereas medical risk was particularly high in Match 2 and Mismatch 1, comprising women who actually delivered by elective CS.
Regarding post-traumatic stress symptoms, we found a significant interaction effect between preference and actual mode of delivery (F = 7.15, p = .008), and in examining differences for all four match/mismatch groups, ANOVA yielded significant overall group differences (F = 11.96, p < 0.001). However, although the Mismatch 1 and Mismatch 2 groups had a tendency toward higher levels of post-traumatic stress symptoms than the Match 1 and Match 2 groups, Bonferroni post-hoc tests found significant differences only between Match 1 (no preference for CS, no elective CS) and Mismatch 2 (preference for CS, no elective CS) ( Table 1 ).
The correlations with putative risk factors showed that post-traumatic stress symptoms were most strongly related to symptoms of depression and anxiety, but fear of childbirth and neuroticism also showed considerable correlations with such symptoms ( Table 2 ). When risk factors were included one by one as covariates in ANCOVA, the inclusion of fear of childbirth had the greatest potency in explaining group differences in post-traumatic stress symptoms, as the F-value of the group effect was reduced from 11.96 to 4.29. Symptoms of depression and anxiety (reduced to 7.35 and 7.61), and to some degree neuroticism and prior PTSD (reduced to 8.46 and 8.79), also decreased the F-value. However, even though the F-value was substantially reduced by some of the covariates, the overall group effect and the difference in the level of post-traumatic stress symptoms between Match 1 and Mismatch 2 still remained significant, as indicated by post hoc tests. When all risk factors were entered blockwise, all risk factors remained statistically significant ( Table 3 ). However, the block with the psychological risk factors showed the greatest potency in explaining group differences, compared to personality variables and somatic and demographic factors. Including the psychological factors reduced the F-value to 3.17, but the group differences remained significant ( Table 3 ). Within the block with psychological risk factors, fear of childbirth was the most important covariate (F = 32.76,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.019), followed by symptoms of depression (F = 26.05,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.015) and anxiety (F = 22.77,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.013). Within the block comprising personality, neuroticism was the most important covariate (F = 78.73,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.044), and within the block with somatic and demographic factors, parity had the largest effect (F = 20.41,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.012) ( Table 3 ).
When all covariates were entered simultaneously in one ANCOVA (they were all statistically significant in the previous model and thus included), the F-value was further reduced to 2.66; however, the group differences remained significant ( Table 3 ). In this model, symptoms of depression (F = 26.55,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.016), parity (F = 22.13,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.013), fear of childbirth (F = 20.85,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.012) and symptoms of anxiety (F = 17.74,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.010) were the most important covariates. The other risk factors (except prior PTSD, F = 3.98,
?dctmLink chronic_id='0901c79180790183' object_id='0901c79180790183' edit_widget_type=graphic??dctmEditor selectedObject='0901c79180790183'? = 0.002) were no longer statistically significant ( Table 3 ).
All ANCOVA were, furthermore, repeated with log transformed scores of the Impact of Event Scale. Besides minor changes in the F-values, the results were very similar to the initial analyses without transformation.